What a negative prompt can and cannot remove

The mechanism, in one paragraph

Classifier-free guidance compares a conditioned prediction against an unconditional one. A negative prompt replaces “unconditional” with “conditioned on this other text”, and the model is pushed away from it:

prediction = neg + scale * (pos - neg)

So a negative prompt does not delete anything. It moves the target away from a region. Whether that helps depends entirely on whether the thing you named is a direction in that space or an object in the picture.

What works: qualities of rendering

Terms describing how the image looks rather than what is in it:

• blurry, low quality, compression artefacts, jpeg artifacts
• warping, morphing, distorted, deformed
• extra limbs, extra fingers, fused fingers
• watermark, text overlay, logo, subtitles, timestamp
• oversaturated, washed out, flat lighting
• flickering, strobing, stuttering

These work because they name a broad region of bad output that the model has seen a great deal of. Pushing away from that region improves things across the frame.

The video-specific ones (warping, morphing, flickering) are worth having in a default negative, because they describe exactly the failures video models produce.

What fails: objects

Put no dog or dog in a negative prompt and you will often get a dog.

The reason is worth understanding. Text conditioning does not carry negation well; the token for the concept activates the concept. In the positive prompt, saying “a street with no dogs” reliably produces dogs, for the same reason telling somebody not to think of an elephant does not work. In the negative prompt the subtraction is real but weak, because one noun against a whole conditioned prediction is a small vector.

What happens in practice is that the concept gets activated and only partly subtracted, so you get a dog-shaped smear, or a dog in the background, or a dog-adjacent animal.

The reliable way to remove an object from a frame is to remove it from the frame. Repair the still. Clone it out. This is the point of the repair lesson and it applies with full force here.

Negatives can suppress motion you want

A trap specific to video. A negative list containing motion blur and movement and camera shake will produce a stiffer clip than you intended, because you have pushed away from the region of the space that contains movement generally.

If your clips come back oddly frozen, read your negative prompt before you touch anything else. This is a common and quietly expensive mistake.

Keep the negative list short and about rendering failures. Four to eight terms. Long copied-and-pasted negative lists from image-model forums are a liability here.

Distilled models often ignore it entirely

A model trained to work at guidance 1 is not doing the extrapolation the formula above describes, so the negative branch may have little or no effect. Some interfaces still show the box.

Test it once: generate with an absurd negative prompt (blue, sky, person) and see whether anything changes. If nothing does, stop typing in that box and save yourself the time.

Prompt weighting, briefly

Some tools accept emphasis syntax such as (wet pavement:1.3) to push a term harder. Where it exists it is worth knowing, and it obeys the same rule as everything else: it steers emphasis, it does not supply information the model lacks. Weighting (logo:1.5) does not make the model know your logo.

Overuse produces the same burnt, frozen output as excessive guidance, because it is the same operation applied to one token.

The summary

A negative prompt is a small, free improvement for rendering quality and a poor tool for content control. The hierarchy of what actually controls a generated clip, strongest first:

1. The first frame.
2. The conditioning signal, if you supplied one.
3. Camera and motion controls.
4. The positive prompt.
5. The negative prompt.

People spend their time in the reverse order. Spending it in this order is most of what separates a production workflow from an afternoon of prompt tinkering.

Today: shorten your negative prompt to six terms about rendering quality, remove anything that names an object or a kind of movement, and see whether your clips get looser and better.